Paddy McMaster Consultant in Paediatric Intensive Care University Hospital of North Staffordshire Stoke on Trent UK
Haemofiltration Plasmafiltration In sepsis / SIRS Excluded: not sepsis, subgroup analysis unclear
Systemic Inflammatory Response Syndrome Toxins Inflammatory response Remove the evil humors Bacterial toxins Lipo-polysaccharide endotoxin Exotoxins Pro-inflammatory cytokines e.g. TNFa, IL1/2/8, PAF, C3a/5a, etc +/- replace consumed factors e.g. IL6/10, Ig, protein C
Fluids Crystalloid Albumin Plasma Access Arterio venous Veno venous Infective agent route Animal Duration Flow rates Filtration rate Blood flow Additional techniques Dialysis Adsorption Outcome measures Survival Clinical Immunological
Rats, 1 Sheep, 1 Pigs, 7 Dogs, 4
Infective agent Endotoxin Bacterial culture (E. coli, S. aureus, P. aerug) IV, intraperitoneal clot, inhalation Sup. Mesenteric Art / caecal ligation Access: CAVH 8; CVVH 5 Start time: before?bp vs after
Filtration ml/kg/hr Duration hours Mortality treatment Mortality control Other effects of treatment 33 3 0/12 0/26 Improved haemodynamics 50-80 2-3 0/8 0/14 Improved LV contractility 80 3 0/8 0/8 Higher mean blood pressure 60 6 6/7 11/14 None
Filtration ml/kg/hr Duration hours Mortality treatment Mortality control Other effects of treatment 160 3.5 0/6 1/12 Improved haemodynamics 167 2.5 2/6 6/6 Improved haemodynamics 1.2-2.5 6 20/20 20/20 Longer survival 100-200 6 6/7 7/7 Longer survival with larger pore 20 4.5 4/10 7/10 Lung mechanics better 8 24 0/8 0/8 No clinical difference 25 24 0/7 0/14 None
(Rogiers 2006) BP (mmhg) Blood warmed 90 Control 38 Cardiac output (l/min) Mortality <16h 4.0 0/10 2.3 10/10 Metabolic acidosis & Lactate less in warmed
TNF & PGF1? (Bellomo 2000) Not significantly different (Ishihara 1999) Pore size (Lee 1998) 50 KD survival 56 hrs 100 KD survival 103 hrs Ultrafiltrate (Grootendorst 1992) Causes death on infusion to non-septic pigs
Control (Freeman 1995) Not filtered (mortality 14%) Filtered (mortality 29%) +/- ultrafiltrate re-infused (Lee 1993) Outcome 2 studies numbers too small for statistical significance 8 benefit 3 no benefit
Blood flow ml/min Duration hours Mortality Effect of treatment 100-0/1 Improved haemodynamic 450 4 11/20 Improved haemodynamic 300 4 2/24 Improved haemodynamic - 36 1/1 Coagulopathy improved 400-750 - 7/9 None 2-145 26 14/19 None
Conditions Sepsis +/- ARDS (Gotloib 1986) Outcomes ARF (Tonnesen 1993) Cardiac index improved 50% SvO 2 increased 25% Inotropes decreased requirement Others no significant metabolic, resp, haem change Bias in reporting +ve outcomes
9 prospective studies 5 non-randomised cohort : n = 5 (Hoffmann 1999) 10 (Kodama 1997) 16 (Hoffmann 1995) 18 (Heering 1997) 18 (Bellomo 1995) Conditions: Sepsis (6), SIRS (3) +/- MOF (Bellomo 1995) ARF (Bellomo 1993; Heering 1997; Kellum 1998) Filtration rates 0.4 2 L/hr (6-30 ml/kg/hr)
Outcome: 4 cytokine levels (no clinical differences) TNF & IL-1 in ultrafiltrate but levels unchanged (none in controls) 2 effect of ultrafiltrate?myocyte contractility Monocyte?TNF,?IL-2 & IL-6 Cohort mortality rates 7/10 (Kodama 1997) 12/18 (Heering 1997) Non-significant? mortality
Trial design Flow ml/min Duration hrs Mortality Effect RCT 150 48 9/13 Nonsignificant? mortality RCT 100-120 48 5/15 Nonsignificant? mortality RCT (adsorption) - 2 17/37 Nonsignificant? mortality RCT (crossover) - 24 12/13 None
molecular adsorbents recirculating system (MARS) CVVH
Animals Rats (Cohen 1987) Can tolerate endotoxin > humans Rabbits (Tetta 2000) Toxin 109 rabbits in 9 arms Similar sensitivity to LPS of E. coli as humans Endotoxin (3), E. coli (IP Natanson 1993; IV Busund 1991) Adsorption Polymyxin B (binds endotoxin) (Cohen 1987) Reverse phase resin (Tetta 2000)
Haemodynamic improved - - 1.25 30 PF Pigs Haemodynamic worse 5/6 6/6 2 100 PF Dogs Haemodynamic improved 7/8 0/6 1.5 20 PF Dogs No difference haemodynamic 30/40 3/19 3 10 PF & adsorptn Rabbit Haematological improved 4/4 0/4 1.5 - PF & adsorptn Rats Other effects of treatment Mortality control Mortality treatment Duration hours Blood flow ml/min Method Animal
Condition 6 Meningococcal 2 Pneumococcal 2 Sepsis including feb neut, VZV, HUS 3 single cases & 3 series of 3 cases all survived 1 blood exchange (survived) (van Deuren 1992) Mortality: 2/6, 1/8, 6/9, 8/12, 0/3 Haemodynamic 1: Improvement SVRI, CI, LVEJ, DO 2 (Berlot 1997) 1: No difference VI, A-a DO 2, inotropes (Reeves 1995)
Technique Condition treated Mortality treatment Mortality control Difference Plasma exchange Meningococcaemia 1/13 6/10 P=0.025 Meningococcaemia Leukaplasmapheresis 3/13 7/9 P=0.02 Plasma exchange & CVVHF Septic shock 1/7 8/21 P=0.25 Plasmapheresis Surgical sepsis 11/19 13/24 P=0.94
Mortality Control Plasmapheresis Reeves 1999 6/14 8/16 P=0.73 No difference survival or number organs failing Busund 2002 18/52 28/52 P=0.05 28d relative risk mortality 0.61 Number needed to treat 4.9
PCCM 29 May 2007 Publish Ahead of Print Shaheen et al Nottingham, Sheffield, Leicester
Bacterial sepsis 30 CVVH 6 CVVH + ECMO Viral sepsis 2 CVVH 14 CVVH + ECMO All sepsis with MOF 22/56 (39%) survived 4 meningococcal plasmafiltration 2 survived
22.2.06 1y/o? 150 ml/kg before admission Adrenaline 1.6 mcg/kg/min Noradrenaline 1 mcg/kg/min Milrinone 750 ng/kg/min Vasopressin ph 6.97, BE -20 CVVH Blood flow 70 ml/min Replacement 300 ml/hr (30 ml/kg/hr) Balance neutral Anuric after 6hrs
Haemofiltration Animals: 8 benefit vs 3 no benefit Human: 3 case studies improved haemodynamics 12 studies (inc 3 RCT) no? mortality Plasmafitration Animals: 3 improved, 1 worse Human: retrospective mortality lower 2; same 2 prospective mortality lower 1; same 1